Mass dependence and scaling properties of nuclear short- range … · 2013-08-21 · Mass dependence and scaling properties of nuclear short- range correlations Wim Cosyn, Camille

Mass dependence and scaling properties of nuclearshort- range correlations

Wim Cosyn, Camille Colle, Maarten Vanhalst, Jan Ryckebusch

Ghent University, Belgium

JLab Physics Division Seminar16th August, 2013

Outline

1 Introduction

2 Our model

3 Inclusive measurements

4 Exclusive reactions

5 EMC vs SRC

M.V., W.C., J.R., PRC84 031302(R) ('11),PRC86 044619 ('12),

arXiv:1210.6175

Wim Cosyn (UGent) JLab Physics Division Seminar August 16, 2013 2 / 30

Introduction

V(r)

VNN(r)

1fm

I Historically: nuclear physics with hadronicdegrees of freedom

baryons (p, n,..) and mesons (π , ρ,..) nuclear shell model: protons and neutrons

in a average mean-field potentialI 1960s-70s: discovery of quantum chromo-

dynamics (QCD) with partonic d.o.f. How do colourless hadrons, nuclei emerge

from QCD? Partonic properties of hadrons change in a

dense medium (EMC effect)I Mean-field approach has limitations

Nucleus is more than sum of nucleons. hard core of the NN-potential induces

short-range correlations (SRC) induce high density & high momentum

fluctuations in the nucleus, deplete shellmodel levels.


Introduction

V(r)

VNN(r)

1fm











Introduction

VNN(r)

1fm











Momentum distributions

dσ (e,e ′p)A = Kσ epρD (~pm)

I Hard core of the NN potential inducesSRC → high-momentum tails

I Lots of nuclear structure activity incomputing one- and two-bodymomentum distributions using ab initiomethods

I In experiments, one-body and two-bodymomentum distributions are not directlyobservable and the obtained informationon SRC is indirect

I f.i. A(e, e ′p) cross section only factorizesin non-relativistic approximation


Momentum distributions

dσ (e,e ′p)A = Kσ epρD (~pm)

I Hard core of the NN potential inducesSRC → high-momentum tails

I Lots of nuclear structure activity incomputing one- and two-bodymomentum distributions using ab initiomethods

I In experiments, one-body and two-bodymomentum distributions are not directlyobservable and the obtained informationon SRC is indirect

I f.i. A(e, e ′p) cross section only factorizesin non-relativistic approximation


Short-range correlations in nuclei

-0.04

-0.02

0

0 0.5 1 1.5 2 2.5 3 3.5 4

f t τ

(r12

)

r12(fm)

0 0.2 0.4 0.6 0.8

1

0 0.5 1 1.5 2 2.5 3 3.5 4

g c (

r 12)

r12(fm)

GDCBF

PieperCluster

10-7

10-6

10-5

10-4

10-3

0 300 600 900

|f t τ

(k12

)|2 (

fm6 )

k12(MeV)

|ΨD(k12)|2

10-7

10-6

10-5

10-4

10-3

0 300 600 900

|gc(

k 12)

|2 (fm

6 )

k12(MeV)

I Time-honoured method is the use ofcorrelation functions

| Ψ 〉 = 1√〈 ΨMF | G† G | ΨMF 〉

G | ΨMF〉

I Dominated by central and tensorcorrelations: highly local → universalcharacter

I In practice: perturbative (cluster, virial)expansions are required to computecorrelated part of momentum distribution

I Local character of SRC truncatesexpansion (2N SRC 3N SRC)

I Ab initio calculations get very compli-cated for medium and heavy nuclei


Short-range correlations in nuclei

-0.04

-0.02

0

0 0.5 1 1.5 2 2.5 3 3.5 4

f t τ

(r12

)

r12(fm)

0 0.2 0.4 0.6 0.8

1

0 0.5 1 1.5 2 2.5 3 3.5 4

g c (

r 12)

r12(fm)

GDCBF

PieperCluster

10-7

10-6

10-5

10-4

10-3

0 300 600 900

|f t τ

(k12

)|2 (

fm6 )

k12(MeV)

|ΨD(k12)|2

10-7

10-6

10-5

10-4

10-3

0 300 600 900

|gc(

k 12)

|2 (fm

6 )

k12(MeV)

I Time-honoured method is the use ofcorrelation functions

| Ψ 〉 = 1√〈 ΨMF | G† G | ΨMF 〉

G | ΨMF〉

I Dominated by central and tensorcorrelations: highly local → universalcharacter

I In practice: perturbative (cluster, virial)expansions are required to computecorrelated part of momentum distribution

I Local character of SRC truncatesexpansion (2N SRC 3N SRC)

I Ab initio calculations get very compli-cated for medium and heavy nuclei


Response to EW interaction

I One-body operator matrix element 〈Ψ|Ω|Ψ〉 = 〈ΨMF| G† ΩG |ΨMF〉I Effective one-body operator receives two-body etc. contributions

through the correlation operators.

G† Ω G ≈ Ω +A∑

i<j=1

([Ω[1](i ) + Ω[1](j )

] [−gc

(rij)

+ t (i , j )]+ h.c.

).

Ω Ωeff

I Two-nucleon knockout A(e, e ′NN) is the hallmark of SRC inexperiments


Response to EW interaction

I One-body operator matrix element 〈Ψ|Ω|Ψ〉 = 〈ΨMF| G† ΩG |ΨMF〉I Effective one-body operator receives two-body etc. contributions

through the correlation operators.

G† Ω G ≈ Ω +A∑

i<j=1

([Ω[1](i ) + Ω[1](j )

] [−gc

(rij)

+ t (i , j )]+ h.c.

).

Ω Ωeff

I Two-nucleon knockout A(e, e ′NN) is the hallmark of SRC inexperiments


Our Model

Quantify the amount of correlated pairs

Approximate method that covers the whole A-range

I Correlation functions require strength at r12 ≈0

I Harmonic oscillator basis: coordinate transformation~r1~r2

~R~r12

I Analytical basis transformation through Standard Moshinsky Brackets

φn1l1 (~r1)φn2l2 (~r2)

φNL(~R)φnL(~r12)

〈n1l1n2l2|NLnL〉SMB

I φnl(~r ) ∼ rl → Only L = 0 (relative S-wave) has strength at r12 ≈0

Identify n = 0,L = 0 pairs in the mean-field wf as prone to SRC !!


In formulas

I |α1α2; JM〉na =∑

n12l12N12Λ12LMLSMSTMT

j1 j2LS√2(1+δα1α2 )

[1− (−1)l12+S+T ]

×〈n12l12N12Λ12(L) | n1l1n2l2(L)〉SMB

l1 l2 L1

2

1

2S

j1 j2 J

〈LMLSMS | JM〉

×〈12t1

1

2t2 | TMT 〉

∣∣∣[n12l12 (~r12),N12Λ12

(~R12

)]LML, SMS ,TMT

⟩

I Normalization: A(A−1)2

=∑

JMα1α2

na 〈α1α2; JM |α1α2; JM〉na

I Suggestion: number of correlated pairs

N(A,Z ) =∑

JMα1α2

na

⟨α1α2; JM

∣∣∣Pn=0,l12=0

~r12

∣∣∣ α1α2; JM⟩

na


Dominance of n = 0,L = 0 pairs

56Fe two-body relative momentum distribution

10−3

10−2

10−1

100

101

102

103

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

n2n+l

2(k

12)

[GeV

−3]

k12 [GeV]

2n+ l = 02n+ l = 12n+ l ≥ 2

allIPM

I First order cluster expansionI Contribution of different

relative quantum numbersI Clear dominance of IPM

n = 0, l = 0 above the Fermimomentum

I Correlation operators canchange the quantum numbersof course


Strength of correlation functions

Overlap of correlation function with different contributions

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

l = 0 1 2 3 4 5

Al C

entr

al C

orr

Str

ength

n=0n=1n=2

-4

-2

0

2

4

6

8

10

12

14

l = 0 1 2 3 4 5A

l T

ensor

Corr

Str

ength

n=0n=1n=2

-3

-2.5

-2

-1.5

-1

-0.5

0

l = 0 1 2 3 4 5

Fe C

entr

al C

orr

Str

ength

n=0n=1n=2n=3

-10

-5

0

5

10

15

20

25

30

35

l = 0 1 2 3 4 5

Fe T

ensor

Corr

Str

ength

n=0n=1n=2n=3

-12

-10

-8

-6

-4

-2

0

l = 0 1 2 3 4 5

Pb C

entr

al C

orr

Str

ength

n=0n=1n=2n=3n=4n=5

-40

-20

0

20

40

60

80

100

120

140

l = 0 1 2 3 4 5

Pb T

ensor

Corr

Str

ength

n=0n=1n=2n=3n=4n=5

I Integrated effect of the tensor correlations is larger by a factor of ≈ 10compared to central correlations

I n = 0,L = 0 contributions dominate


Distribution of the relative quantum numbersL = S ,P,D, F ,G ,H, I ,≥ J for pp pairs


Distribution of the relative quantum numbersL = S ,P,D, F ,G ,H, I ,≥ J for pn pairs


Distribution of the relative quantum numbersL = S ,P,D, F ,G ,H, I ,≥ J for pn pairs

with increasing A: smaller fraction of the pairs residein a relative S-state

strong isospin dependence: fraction of the pn pairs residing in arelative S-state is substantially larger than for pp and nn pairs.


Number of pp, nn and pn pairs with L = 0

0.1

1

10

100

1000

1 10 100

num

ber

of

S p

air

s

mass number A

np(T=0)HOnp(T=0)WS

np(T=1)ppnnFit

power law ∼ A1.44±0.01

I Very soft A-dependence(naive A2)

I Power law is robust

I Isospin dependence


Inclusive electronscattering measurements

A-dependence of SRC: experiment

0

200

400

600

800

1000

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

|pm

min

| (M

eV)

xB

Q2 = 0.5 GeV2

Q2 = 1.0 GeV2

Q2 = 1.5 GeV2

Q2 = 2.0 GeV2

Q2 = 2.5 GeV2

Q2 = 3.0 GeV2

Q2 = 3.5 GeV2

Q2 = 4.0 GeV2

I Inclusive A(e,e’) scattering atBjorken x > 1.4 and high Q2

I SRC universality: Cross sectionratios to the deuteron showscaling σA = a2

A2σD → a2 is

measure for the amount ofcorrelated pairs

I Compared to deuteron correlatedpair in nucleus A also has

Binding energy Center of mass motion FSI with nuclear

medium



3

at large x, where scattering from nucleons below theFermi momentum is forbidden. If these high-momentumcomponents are related to two-nucleon correlations (2N-SRCs), then they should yield the same high-momentumtail whether in a heavy nucleus or a deuteron.

The first detailed study of SRCs in inclusive scatteringcombined data from several measurements at SLAC [12],so the cross sections had to be interpolated to identicalkinematics to form the ratios. A plateau was seen in theratio (σA/A)/(σD/2) that was roughly A-independent forA ≥ 12, but smaller for 3He and 4He. Ratios from Hall Bat JLab showed similar plateaus [13, 14] and mapped outthe Q2 dependence at lowQ2, seeing a clear breakdown ofthe picture for Q2 < 1.4 GeV2. However, these measure-ments did not include deuterium; only A/3He ratios wereavailable. Finally, JLab Hall C data at 4 GeV [15, 16]measured scattering from nuclei and deuterium at largerQ2 values than the previous measurements, but the deu-terium cross sections had limited x coverage. Thus, whilethere is significant evidence for the presence of SRCsin inclusive scattering, clean and precise ratio measure-ments for a range of nuclei are lacking.

0

3

6 3He

0

3

64He

(σA/A

)/(σ

D/2

)

0

3

6

0.8 1 1.2 1.4 1.6 1.8

x

9Be

12C

63Cu

1 1.2 1.4 1.6 1.8 2

x

197Au

FIG. 2: Per-nucleon cross section ratios vs x at θ=18.

Figure 2 shows the A/D cross section ratios for theE02-019 data at a scattering angle of 18. For x > 1.5,the data show the expected near-constant behavior, al-though the point at x = 1.95 is always high because the2H cross section approaches zero as x → MD/Mp ≈ 2.This was not observed before, as the previous SLAC ra-tios had much wider x bins and larger statistical uncer-tainties, while the CLAS took ratios to 3He.

Table I shows the ratio in the plateau region for a rangeof nuclei at all Q2 values where there was sufficient large-x data. We apply a cut in x to isolate the plateau region,although the onset of scaling in x varies somewhat withQ2. The start of the plateau corresponds to a fixed valueof the light-cone momentum fraction of the struck nu-cleon, αi [1, 12]. However, αi requires knowledge of the

initial energy and momentum of the struck nucleon, andso is not directly measured in inclusive scattering. Thus,the plateau region is typically examined as a function ofx or α2n, which corresponds to αi under the approxi-mation that the photon is absorbed by a single nucleonfrom a pair of nucleons with zero net momentum [12]. Wetake the A/D ratio for xmin < x < 1.9, such that xmin

corresponds to a fixed value of α2n. The upper limit isincluded to avoid the deuteron kinematic threshold.

TABLE I: r(A,D) = (2/A)σA/σD in the 2N correlation re-gion (xmin < x < 1.9). We choose a conservative value ofxmin = 1.5 at 18, which corresponds to α2n = 1.275. We usethis value to determine the xmin cuts for the other angles.The last column is the ratio at 18 after the subtraction ofthe estimated inelastic contribution (with a systematic uncer-tainty of 100% of the subtraction).

A θ=18 θ=22 θ=26 Inel.sub3He 2.14±0.04 2.28±0.06 2.33±0.10 2.13±0.044He 3.66±0.07 3.94±0.09 3.89±0.13 3.60±0.10Be 4.00±0.08 4.21±0.09 4.28±0.14 3.91±0.12C 4.88±0.10 5.28±0.12 5.14±0.17 4.75±0.16Cu 5.37±0.11 5.79±0.13 5.71±0.19 5.21±0.20Au 5.34±0.11 5.70±0.14 5.76±0.20 5.16±0.22〈Q2〉 2.7 GeV2 3.8 GeV2 4.8 GeV2

xmin 1.5 1.45 1.4

At these high Q2 values, there is some inelastic contri-bution to the cross section, even at these large x values.Our cross section models predicts that this is approxi-mately a 1–3% contribution at 18, but can be 5–10% atthe larger angles. This provides a qualitative explanationfor the systematic 5–7% difference between the lowest Q2

data set and the higher Q2 values. Thus, we use only the18 data, corrected for our estimated inelastic contribu-tion, in extracting the contribution of SRCs.The typical assumption for this kinematic regime is

that the FSIs in the high-x region come only from rescat-tering between the nucleons in the initial-state correla-tion, and so the FSIs cancel out in taking the ratios [1–3, 12]. However, it has been argued that while the ratiosare a signature of SRCs, they cannot be used to providea quantitative measurement since different targets mayhave different FSIs [17]. With the higher Q2 reach ofthese data, we see little Q2 dependence, which appearsto be consistent with inelastic contributions, supportingthe assumption of cancellation of FSIs in the ratios. Up-dated calculations for both deuterium and heavier nucleiare underway to further examine the question of FSI con-tributions to the ratios [18].Assuming the high-momentum contribution comes en-

tirely from quasielastic scattering from a nucleon in ann–p SRC at rest, the cross section ratio σA/σD yieldsthe number of nucleons in high-relative momentum pairsrelative to the deuteron and r(A,D) represents the rela-tive probability for a nucleon in nucleus A to be in such

data: Fomin et al. (JLab Hall C), PRL108 092502



A2σD → a2 is




medium



3

at large x, where scattering from nucleons below theFermi momentum is forbidden. If these high-momentumcomponents are related to two-nucleon correlations (2N-SRCs), then they should yield the same high-momentumtail whether in a heavy nucleus or a deuteron.

The first detailed study of SRCs in inclusive scatteringcombined data from several measurements at SLAC [12],so the cross sections had to be interpolated to identicalkinematics to form the ratios. A plateau was seen in theratio (σA/A)/(σD/2) that was roughly A-independent forA ≥ 12, but smaller for 3He and 4He. Ratios from Hall Bat JLab showed similar plateaus [13, 14] and mapped outthe Q2 dependence at lowQ2, seeing a clear breakdown ofthe picture for Q2 < 1.4 GeV2. However, these measure-ments did not include deuterium; only A/3He ratios wereavailable. Finally, JLab Hall C data at 4 GeV [15, 16]measured scattering from nuclei and deuterium at largerQ2 values than the previous measurements, but the deu-terium cross sections had limited x coverage. Thus, whilethere is significant evidence for the presence of SRCsin inclusive scattering, clean and precise ratio measure-ments for a range of nuclei are lacking.

0

3

6 3He

0

3

64He

(σA/A

)/(σ

D/2

)

0

3

6

0.8 1 1.2 1.4 1.6 1.8

x

9Be

12C

63Cu

1 1.2 1.4 1.6 1.8 2

x

197Au

FIG. 2: Per-nucleon cross section ratios vs x at θ=18.

Figure 2 shows the A/D cross section ratios for theE02-019 data at a scattering angle of 18. For x > 1.5,the data show the expected near-constant behavior, al-though the point at x = 1.95 is always high because the2H cross section approaches zero as x → MD/Mp ≈ 2.This was not observed before, as the previous SLAC ra-tios had much wider x bins and larger statistical uncer-tainties, while the CLAS took ratios to 3He.

Table I shows the ratio in the plateau region for a rangeof nuclei at all Q2 values where there was sufficient large-x data. We apply a cut in x to isolate the plateau region,although the onset of scaling in x varies somewhat withQ2. The start of the plateau corresponds to a fixed valueof the light-cone momentum fraction of the struck nu-cleon, αi [1, 12]. However, αi requires knowledge of the

initial energy and momentum of the struck nucleon, andso is not directly measured in inclusive scattering. Thus,the plateau region is typically examined as a function ofx or α2n, which corresponds to αi under the approxi-mation that the photon is absorbed by a single nucleonfrom a pair of nucleons with zero net momentum [12]. Wetake the A/D ratio for xmin < x < 1.9, such that xmin

corresponds to a fixed value of α2n. The upper limit isincluded to avoid the deuteron kinematic threshold.

TABLE I: r(A,D) = (2/A)σA/σD in the 2N correlation re-gion (xmin < x < 1.9). We choose a conservative value ofxmin = 1.5 at 18, which corresponds to α2n = 1.275. We usethis value to determine the xmin cuts for the other angles.The last column is the ratio at 18 after the subtraction ofthe estimated inelastic contribution (with a systematic uncer-tainty of 100% of the subtraction).

A θ=18 θ=22 θ=26 Inel.sub3He 2.14±0.04 2.28±0.06 2.33±0.10 2.13±0.044He 3.66±0.07 3.94±0.09 3.89±0.13 3.60±0.10Be 4.00±0.08 4.21±0.09 4.28±0.14 3.91±0.12C 4.88±0.10 5.28±0.12 5.14±0.17 4.75±0.16Cu 5.37±0.11 5.79±0.13 5.71±0.19 5.21±0.20Au 5.34±0.11 5.70±0.14 5.76±0.20 5.16±0.22〈Q2〉 2.7 GeV2 3.8 GeV2 4.8 GeV2

xmin 1.5 1.45 1.4

At these high Q2 values, there is some inelastic contri-bution to the cross section, even at these large x values.Our cross section models predicts that this is approxi-mately a 1–3% contribution at 18, but can be 5–10% atthe larger angles. This provides a qualitative explanationfor the systematic 5–7% difference between the lowest Q2

data set and the higher Q2 values. Thus, we use only the18 data, corrected for our estimated inelastic contribu-tion, in extracting the contribution of SRCs.The typical assumption for this kinematic regime is

that the FSIs in the high-x region come only from rescat-tering between the nucleons in the initial-state correla-tion, and so the FSIs cancel out in taking the ratios [1–3, 12]. However, it has been argued that while the ratiosare a signature of SRCs, they cannot be used to providea quantitative measurement since different targets mayhave different FSIs [17]. With the higher Q2 reach ofthese data, we see little Q2 dependence, which appearsto be consistent with inelastic contributions, supportingthe assumption of cancellation of FSIs in the ratios. Up-dated calculations for both deuterium and heavier nucleiare underway to further examine the question of FSI con-tributions to the ratios [18].Assuming the high-momentum contribution comes en-

tirely from quasielastic scattering from a nucleon in ann–p SRC at rest, the cross section ratio σA/σD yieldsthe number of nucleons in high-relative momentum pairsrelative to the deuteron and r(A,D) represents the rela-tive probability for a nucleon in nucleus A to be in such

data: Fomin et al. (JLab Hall C), PRL108 092502



A2σD → a2 is




medium


Nuclear corrections for a2Effect of A− 2 excitation energy

MC simulations of breakup of 2N correlatedpairs in 12C for ε = 5.766 GeV and⟨

Q2⟩=2.7 GeV2

Effect of c.m. motion of pn pairs


Nuclear corrections for a2Effect of A− 2 excitation energy

MC simulations of breakup of 2N correlatedpairs in 12C for ε = 5.766 GeV and⟨

Q2⟩=2.7 GeV2

Effect of c.m. motion of pn pairs


Mass dependence: comparison experiment - calculations

0

1

2

3

4

5

6

7

8

1 10 100

a 2 (

A/d

)

mass number A

Data: K. Egiyan et al. (CLAS), PRL96 082501 ('06),N. Fomin et al. (JLab Hall C), PRL108 092502 ('12)

L. Frankfurt et al. (SLAC), PRC48 2451 ('93)

I Assumed dominance of np pairswith relative n = 0, l = 0(S = 1)

I a2 ≈2

ANpn(S=1)(A,Z )

∫PS

d ~P12Fpn (P12)

I Corrections of the c.m. motionand binding applied

I FSI being worked on...I Prediction: a2(40Ca) ≈ a2(48Ca )

[preliminary Hall A E08014 dataseems to agree]

I Missing strength at low A due toclustering?

I Overestimation at high A.I Mass dependence much softer

than NZ .


Can one quantify the number of 3N correlations?

Three-body correlations induced by tensor correlations(ftτ (r12)S12

) [fLS (r23)~L23 · ~S23

]9

3 3p

p

n

S=1, T=0, L=0

S=0, T=1, L=0

pp

n

S=1, T=0, L=2

S=1, T=1, L=1

uncorrelated correlated

11

2 22

FIG. 12. (Color online). Illustration of three-body correlations in-duced by tensor correlations. In the uncorrelated wave function (left)the two nucleons 1 and 2 are in anS=1, MS=0 pair with L=0. Thetensor force leads to an admixture of anL=2 component and an align-ment of the spins of nucleons 1 and 2 flipping the spin of nucleon 2(right). This affects the interaction between nucleon 2 and nucleon 3.In the uncorrelated wave function the protons 2 and 3 form anS=0,T=1, L=0 pair. After the spin-flip of nucleon 2 this becomes anS=1,T=1, L=1 pair.

10-5

10-4

10-3

10-2

10-1

100

101

102

0 1 2 3 4

nrel

S,T

(k)

[fm

3 ]

k [fm-1]

(ST)=(01)

(ST)=(11)

T=1

FIG. 13. (Color online). Two-body densities in momentum space for4He in theS=0, T=1 andS=1, T=1 channels and the sum of bothdensities.

that the two-body densities in theS=1,T=0 channel are muchlarger than in theS=0,T=1 channel in this momentum region.

As already mentioned this effect can not be understood interms of two-body correlations. It also explains why effectiveinteractions that are obtained by unitary transformationsintwo-body approximation, likeVlow−k [35], the similarity renor-malization group (SRG) [36] or the unitary correlation opera-tor method (UCOM) [37–39], provide more binding than thebare interaction when used in exact calculations. In two-bodyapproximation the interaction is transformed independently inall spin-isospin channels. It is therefore possible to obtain thefull contribution of the tensor force in theS=1, T=0 chan-nel without having to pay the price of the three-body corre-lations. With increasing range of the tensor correlations (inthe UCOM approach) or a lower cut-off (in theVlow−k or SRGapproaches) the effective interaction will induce smaller three-body correlations. Smaller three-body correlations meansthatless nucleon pairs are moved from theS=0, T=1 to theS=1,T=1 channel. As in the odd channel the potential is less at-

tractive and the kinetic energy is much larger, the three-bodycorrelations provide a repulsive contribution to the energy.

In the early seventies it has already been realized that a termin the effective interaction called antisymmetric spin-orbit(ALS)-force that connectsS=0 with S=1 states and changesthe relative angular momentum by∆L=1, like (l1 − l2) · (σ1 −σ2), is able to improve spectra and transition rates insd-shellmodel calculations [40]. But as such a term is not conservingtranslational and Galilei invariance it is not allowed in the freenucleon-nucleon interaction and can only be obtained by inte-grating many-body forces over additional particle degreesoffreedom.

We want to stress the point that in our discussion no gen-uine three-body forces are considered. The three-body cor-relations are induced by the two-body tensor force. Whengenuine three-body forces are included we of course expectadditional or modified three-body correlations.

F. Comparison with UCOM

The universality of short-range correlations is not only in-teresting in itself but also confirms the basic assumptions thatunderlie methods to derive effective low-momentum interac-tions like UCOM,Vlow−k and SRG. We will discuss here theUCOM approach as it provides the most direct connection tothe short-range correlations in the nucleus.

The basic idea of the UCOM approach is to imprint theshort-range central and tensor correlations into the nuclearmany-body wave functions explicitly by means of unitary cor-relation operatorC. Starting from an uncorrelated trial state|Φ〉 the correlated state

|Ψ〉 = C |Φ〉 (12)

then features the short-range central and tensor correlations.Long-range correlations still have to be incorporated explic-itly in the trial state|Φ〉.

To explain the action of the correlation operators we discussfirst how the relative motion of two nucleons is affected by thecorrelation operators. For that we use basis states

|φ(LS)JM; T MT〉 , (13)

where the relative orbital angular momentumL is coupledwith the spinS of the two nucleons to total angular momen-tum J,M. The isospin is coupled toT,MT . The radial part ofthe relative wave function is given byφ(r).

In the S=0 channels only the central correlation operatoracts and the correlated relative wave function is given, usingthe correlation functionR−(r), as

ψS JTL (r) = 〈r(LS)JT| Cr |φ(LS)JT〉

=R−(r)

r

√R′−(r) φ(R−(r)) ,

(14)

whereas in theS=1 channels both central and tensor corre-lation operators act and we obtain the correlated radial wave

Feldmeier, Horiuchi, Neff, Suzuki: PRC84 054003


Mass dependence of number of ppn triples with(l12 = 0, l(12)3 = 0

)

1

10

100

1000

1 10 100

Num

ber

of p

pn tr

iple

ts

A

0.28 A1.61

ppn

I Number of ppn triples prone to SRC effects (l12 = 0, l(12)3 = 0) :Nppn(A) = 0.28A1.61

I Again very soft A-dependenceWim Cosyn (UGent) JLab Physics Division Seminar August 16, 2013 17 / 30

A(e, e ′) for 2.2 . xB and 3N SRCs

Scaling of the A(e, e ′) response to the3He one

would demonstrate the presence of 3-nucleon (3N) SRC

and confirm the previous observation of NN SRC.

Note that: (i) Refs. [5,6] argue that the c.m. motion of the

NN SRC may change the value of a2 (by up to 20% for56Fe) but not the scaling at xB < 2. For 3N SRC there are

no estimates of the effects of c.m. motion. (ii) Final state

interactions (FSI) are dominated by the interaction of the

struck nucleon with the other nucleons in the SRC [7,8].

Hence the FSI can modify j, while such modification of

ajA are small since the pp, pn, and nn cross sections at

Q2 > 1 GeV2 are similar in magnitudes.

In our previous work [6] we showed that the ratios

RA; 3He 3AQ2;xB

A3HeQ2;xB

scale for 1:5< xB < 2 and 1:4<

Q2 < 2:6 GeV2, confirming findings in Ref. [7]. Here we

repeat our previous measurement with higher statistics

which allows us to estimate the absolute per-nucleon prob-

abilities of NN SRC.

We also search for the even more elusive 3N SRC,

correlations which originate from both short-range NN

interactions and three-nucleon forces, using the ratio

RA; 3He at 2< xB 3.

Two sets of measurements were performed at the

Thomas Jefferson National Accelerator Facility in 1999

and 2002. The 1999 measurements used 4.461 GeV elec-

trons incident on liquid 3He, 4He and solid 12

C targets. The

2002 measurements used 4.471 GeVelectrons incident on a

solid 56Fe target and 4.703 GeV electrons incident on a

liquid 3He target.

Scattered electrons were detected in the CLAS spec-

trometer [9]. The lead-scintillator electromagnetic calo-

rimeter provided the electron trigger and was used to

identify electrons in the analysis. Vertex cuts were used

to eliminate the target walls. The estimated remaining

contribution from the two Al 15 m target cell windows

is less than 0.1%. Software fiducial cuts were used to

exclude regions of nonuniform detector response. Kine-

matic corrections were applied to compensate for drift

chamber misalignments and magnetic field uncertainties.

We used the GEANT-based CLAS simulation, GSIM, to

determine the electron acceptance correction factors, tak-

ing into account ‘‘bad’’ or ‘‘dead’’ hardware channels in

various components of CLAS. The measured acceptance-

corrected, normalized inclusive electron yields on 3He,4He, 12C, and 56Fe at 1< xB < 2 agree with Sargsian’s

radiated cross sections [10] that were tuned on SLAC data

[11] and describe reasonably well the Jefferson Lab Hall C

[12] data.

We constructed the ratios of inclusive cross sections as a

function of Q2 and xB, with corrections for the CLAS

acceptance and for the elementary electron-nucleon cross

sections:

rA; 3He A2ep en

3Zep Nen

3YA

AY3HeRArad; (2)

where Z and N are the number of protons and neutrons in

nucleus A, eN is the electron-nucleon cross section, Y is

the normalized yield in a given (Q2; xB) bin, and RArad

is the

ratio of the radiative correction factors for 3He and nucleus

A [see Ref. [8] ]. In our Q2 range, the elementary cross

section correction factorA2epen

3ZepNenis 1:14 0:02 for C

and 4He and 1:18 0:02 for 56

Fe. Note that the 3He yield

in Eq. (2) is also corrected for the beam energy difference

by the difference in the Mott cross sections. The corrected3He cross sections at the two energies agree within 3:5%

[8].

We calculated the radiative correction factors for the

reaction Ae; e0 at xB < 2 using Sargsian’s upgraded

code of Ref. [13] and the formalism of Mo and Tsai [14].

These factors change 10%–15% with xB for 1< xB < 2.

However, their ratios, RArad

, for 3He to the other nuclei are

almost constant (within 2%–3%) for xB > 1:4. We applied

RArad

in Eq. (2) event by event for 0:8< xB < 2. Since there

are no theoretical cross section calculations at xB > 2, we

applied the value of RArad

averaged over 1:4< xB < 2 to the

entire 2< xB < 3 range. Since the xB dependence of RArad

for 4He and 12

C are very small, this should not affect the

ratio r of Eq. (2). For 56Fe, due to the observed small slope

of RArad

with xB, rA; 3He can increase up to 4% at xB

2:55. This was included in the systematic errors.

Figure 1 shows the resulting ratios integrated over 1:4<

Q2 < 2:6 GeV2. These cross section ratios (a) scale ini-

tially for 1:5< xB < 2, which indicates that NN SRCs

a)

r(4H

e/3

He)

b)

r(1

2C

/3H

e)

xB

r(5

6F

e/3

He)

c)

1

1.5

2

2.5

3

1

2

3

4

2

4

6

1 1.25 1.5 1.75 2 2.25 2.5 2.75

FIG. 1. Weighted cross section ratios [see Eq. (2)] of (a) 4He,

(b) 12C, and (c) 56

Fe to 3He as a function of xB for Q2 >

1:4 GeV2. The horizontal dashed lines indicate the NN (1:5<

xB < 2) and 3N (xB > 2:25) scaling regions.

PRL 96, 082501 (2006)P H Y S I C A L R E V I E W L E T T E R S week ending

3 MARCH 2006

082501-3

JLab Hall-B, PRL96, 082501

I Quantify scaling behavior:

a3(A/3He

)≡ 3

A

σA(xB ,Q2

)

σ 3He(xB ,Q2

) ,

I Assume that signal is dominatedby the ppn correlations!

I Assume thatσeppn(Q2, xB ) ≈ σe3He (Q2, xB )

I Very naive counting (all ppntriples contribute): a3 ∼ A2

I Suggestion:a3(A/3He) ∼ 3

ANppn(A) (number

of ppn triples withl12 = 0, l(12)3 = 0)




4

a configuration. To extract the relative contribution of2N-SRCs, we use the inelastic-subtracted ratios and ap-ply a correction for the smearing effect of the center-of-mass motion of the 2N-SRC pairs. The momentumdistribution of a nucleon in the 2N-SRC will be a con-volution of the relative distribution and the CM motionof the pair, as discussed (and extracted for Carbon) inRef. [19]. This smearing of the distribution enhances thehigh-momentum tails in heavy nuclei, estimated to yielda 20% enhancement for Fe [20]. Thus, for iron we removethis 20% enhancement to the ratio, scale this correctionto the other nuclei based on an estimate of the pair mo-tion as a function of A. We apply an uncertainty equalto 30% of this correction (50% for 3He). In addition toenhancing the high-momentum tails, this effect can alsoyield some distortion in the shape at the largest x values.This may explain the small x dependence in the ratios inFig. 2, which is larger for heavy nuclei.

TABLE II: Extracted value of R2N (A). The results fromSLAC [12] and CLAS [14] have been updated to be consis-tent with the new extraction except for the lack of Coulombcorrection or inelastic subtraction (see text for details).

A R2N (E02-019) SLAC CLAS CM corr3He 1.93±0.10 1.8±0.3 – 1.10±0.054He 3.02±0.17 2.8±0.4 2.80±0.28 1.19±0.06Be 3.37±0.17 – – 1.16±0.05C 4.00±0.24 4.2±0.5 3.50±0.35 1.19±0.06

Cu(Fe) 4.33±0.28 (4.3±0.8) (3.90±0.37) 1.20±0.06Au 4.26±0.29 4.0±0.6 – 1.21±0.06〈Q2〉 ∼2.7 GeV2 ∼1.2 GeV2 ∼2 GeV2

xmin 1.5 – 1.5αmin 1.275 1.25 1.22–1.26

After correcting the measured ratios for the inelas-tic contribution and the enhancement of the high-momentum tails due to motion of the pair, we obtainR2N , given in Tab. II, which represents the probabilty ofa nucleon in nucleus A to be in a high relative momentumpair compared to a nucleon in the deuteron. It also showsupdated extractions from previous data after applyingC.M. motion corrections and removing the “isoscalar”correction factors applied in the previous works. Thiscorrection was based on the assumption that the high-momentum tails would have greater neutron contribu-tions for N>Z nuclei. However, the dominance of isos-inglet pairs [19, 21, 22] implies that the high-momentumtail will have equal proton and neutron contributions.After making consistent extractions from all of the ex-periments, we find a systematic difference between ourdata and the CLAS results. While the CLAS data arecloser to our Q2 value than the SLAC data, their cuton x > 1.5 corresponds to lower minimum value α2n.If α2n is not high enough to fully isolate 2N-SRCs, oneexpects the extracted ratio to be smaller. Our resultshave smaller uncertainties and a more conservative αmin

cut, providing a cleaner extraction of the SRC contribu-tions. Note that the previous experiments do not includeany corrections (or uncertainties) associated with inelas-tic contributions or Coulomb distortion. The latter isestimated to be up to 6% for the Fe data of Ref. [14],and similar or larger for the lower Q2 SLAC data [12].

Previous extractions of the strength of 2N-SRCs founda slow increase of R2N with A in light nuclei, with littleapparent A dependence for A>12. The additional correc-tions applied in our extraction of 2N-SRC contributionsdo not modify these basic conclusions, but these correc-tions, along with the improved precision in our extrac-tion yields a more detailed picture of the A dependence.In a mean-field model, one would expect the probabilityfor two nucleons to be close enough together to form anSRC to be proportional to the average density of the nu-cleus [2]. However, while the density of 9Be is well belowthat of 4He or 12C, and very similar to that of 3He, therelative contribution from SRCs in 9Be is much closer tothat in 4He or 12C, suggesting that the simple expecta-tion that SRCs will scale with density is insufficient. Thisis very much like the recently observed A dependence ofthe EMC effect [23], which also observed that 9Be be-haved more like denser nuclei due to the significant clus-ter structure in 9Be. It seems clear that cluster structureshould be extremely important in examining the short-range structure and contribution of SRCs in nuclei, butit has not been observed before, as previously-measurednuclei did not have sufficient clustering to yield a devia-tion from a simple scaling with density [14].

0

1

2

3

4

5

1 1.5 2 2.5 3

x

(σA/A

)/(σ

He3

/3)

R(4He/3He)

CLASE02-019

FIG. 3: (Color online) The 4He/3He ratios from CLAS andE02-019. Errors shown are the combined statistical and sys-tematic uncertainties. The error bars shown for x ≥ 2.4 (hol-low points) represent the central 68% confidence level region.

For A/3He ratios above x = 2, one expects the two-nucleon contributions to become small so that 3N-SRCsshould eventually dominate. This region was examined inthe CLAS analysis [14], but the statistics did not allowfor an examination of the onset of scaling at high Q2.They assumed that Q2 = 1.4 GeV2 would be sufficientto cleanly isolate 3N-SRCs, but it is not clear that this is

JLab Hall-C, PRL108, 092502

Discrepancy in the data (see alsoprelim. Hall A E08014 results)


a3(A/3He

)≡ 3

A

σA(xB ,Q2

)

σ 3He(xB ,Q2

) ,





ANppn(A) (number





0

2

4

6

8

10

12

14

10 100

Rel

ati

ve p

rob

ab

ilit

y fo

r p

pn

SR

C

mass number A

ppn CLAS data

a3(A/3He) as a measure of theper-nucleon probability of ppn SRCrelative to 3He (calculations are NOTcorrected for c.m. motion, FSI, . . .)


a3(A/3He

)≡ 3

A

σA(xB ,Q2

)

σ 3He(xB ,Q2

) ,





ANppn(A) (number



Exclusive A(e, e ′NN)scattering measurements

Exclusive A(e, e ′pp) reactions

Probe the content of nuclear SRC? → exclusive 2N knockout!!

The fact that SRC-prone proton-proton pairs are mostly in a state withrelative orbital momentum l12 = 0 has important consequences for theEXCLUSIVE A(e, e ′pp) cross sections (J.R. PLB 383,1 (’96))!!

I The A(e, e ′pp) cross sections factorizes according to

d8σdε′dΩε′dΩ1dΩ2dTp2

(e, e ′pp) ≈ KσeN1N2(k+, k−, q)FD (P)

I F (P): probability to find a diproton with c.m. momentum P andrelative orbital momentum l12 = 0!


Factorization of the A(e, e ′pp) cross sections: Experiments!

12C(e, e ′pp) @ MAMI (Mainz)(PLB 421 (1998) 71.)

I Up to P = 0.5 GeV c.o.m.motion in 12C is mean-field(Gaussian) like

I Data agree with thefactorization in terms of F (P)!

I Largest at Pm = 0 →back-to-back


What do experiments say (2)

Triple coincidence measurements A(e, e ′pp) at low Q2 determined thequantum number of the correlated pairs!

16O(e,e′pp)14C(0+; Ex=0.0 MeV)

pair c.m. momentum (MeV/c)

d8 σ/dΩ

e′de

′dΩ

pdΩ

pdT

p (p

b/sr

3 /MeV

2 )

∆SRC + ∆

10-3

10-2

10-1

1

-100 0 100 200 300 400

Unfactorized theory (MEC, IC, central + tensor correlations)EPJA 20 (2004) 435

I High resolution16O(e, e ′pp) studies(MAMI, NIKHEF)

I Ground-state transition:16O (0+)→14C (0+)

I Quantum numbers of theactive diproton [relative(c.m.)]: 1S0(Λ = 0) (lowerP) and 3P1(Λ = 1)(higher P)

I only 1S0(Λ = 0)diprotons are subject toSRC

What do experiments say (3)

2N correlations in 12C(p, 2p + n) at BNL

Tang et al., PRL90, 042301 ('03)Wim Cosyn (UGent) JLab Physics Division Seminar August 16, 2013 22 / 30

What do experiment say (4)

2N correlations in 12C(e, e ′pp) / 12C(e, e ′p) JLAB Hall A

R. Subedi et al., Science 320, 1476 ('08)

R. Shneor et al., PRL99, 072501 ('07)

I 20% of the nucleons arein a SRC pair

I 90% of the correlatedpairs are np pairs →tensor force dominance


The pp c.m. momentum distribution

dσ (e, e ′pp) ≈ KσeN1N2(k+, k−, q)F (P)

50

100

150

200

250

300

P2(P

12)

12C

nl=00 [x5]

all nl

27Al

nl=00 [x5]

all nl

0.0 0.1 0.2 0.3 0.4 0.5P12[GeV]

50

100

150

200

250

300

P2(P

12)

56Fe

nl=00 [x5]

all nl

0.1 0.2 0.3 0.4 0.5P12[GeV]

208Pb

nl=00 [x5]

all nl

I F (P): probability to find adiproton with c.m. momentumP and relative orbitalmomentum l12 = 0!

Widths of c.m. momentumdistributions carry informationabout the quantum numbers of thepairs

I Strength has very softA-dependence


A(e, e ′pp) data from Hall B

xB > 1.2, θpq < 25,

0.62 < p/q < 0.96, pm > 300MeVFig. from Or Hen

I CLAS data mining effort (L.Weinstein et al.)

I Study the influence ofexperimental cuts on F (P) in aMC

I Check: nice agreement withopening angle distribution fromHall A A(e, e ′pp) data

I Widths a little bit (∼ 10 MeV)narrower after the MC

I Influence of FSI checked in ourRMSGA model (relativisticGlauber), widths again narrowby about 10 MeV


A(e, e ′pp) data from Hall BD

istr

ibut

ion

cosγ

12C Uniform12C Gauss

Data

05

1015202530354045

-1 -0.98 -0.96 -0.94 -0.92 -0.9







A(e, e ′pp) data from Hall BD

istr

ibuti

on

P12,x [MeV]

σ = 147 MeV

P12,y [MeV]

σ = 148 MeV

P12,z [MeV]

σ = 129 MeVµ = 93 MeV

|P12| [MeV]

0.0

1.0

2.0

3.0

4.0

-600 -300 0 300 6000.0

1.0

2.0

3.0

4.0

-600 -300 0 300 600

0.0

1.0

2.0

3.0

4.0

-600 -300 0 300 6000.0

1.0

2.0

3.0

4.0

0 300 600







A(e, e ′pp) data from Hall B

0.0 0.2 0.4 0.6 0.8 1.00.0

0.2

0.4

0.6

0.8

1.0

F[f

m−

3 ]

0.00.10.20.30.40.5

12CRPWIA

FSI[×7] 12C

−600 −300 0 300 600P12,x [MeV]

02468

10208Pb

RPWIA

FSI[×30]

−600 −300 0 300 600P12,y [MeV]

208Pb







Center of mass motion of correlated pp pairs

Data is preliminary! Courtesy of CLASdata mining (O. Hen and E. Piasetzky)

I Analysis of exclusiveA(e, e ′pp) for 12C, 27Al, 56Fe,208Pb

I Distribution of events againstP is fairly Gaussian

I σc.m.: Gaussian widths froma fit to measured c.m.distributions

I Theory lines: Gaussian fitsto computed c.m.distributions for l = 0, 1, 2


Mass dependence of the A(e, e ′pp) cross sections

Prediction: A dependence of A(e, e ′pp) cross section is soft(much softer than predicted by naive Z (Z − 1) counting)

A(e, e ′pp)12C(e, e ′pp) ≈

Npp(A)Npp

(12C) ×

(TA(e, e ′p)T12C(e, e ′p)

)1−2

1

10

100

10 100

A(e

,e′p

p)/

12C

(e,e

′pp

)

mass number A

l12=0+FSI

l12=1+FSI

l12=2+FSI


Mass dependence of the A(e, e ′pp) cross sections

Prediction: A dependence of A(e, e ′pp) cross section is soft(much softer than predicted by naive Z (Z − 1) counting)

A(e, e ′pp)12C(e, e ′pp) ≈

Npp(A)Npp

(12C) ×

(TA(e, e ′p)T12C(e, e ′p)

)1−2

1

10

100

10 100

A(e

,e′p

p)/

12C

(e,e

′pp

)

mass number A

l12=0+FSI

CLAS

GLAUBER

I Preliminary data(courtesy of O. Hen andE. Piasetzky) compatiblewith absorption onl12 = 0 pairs!

I Relativized Glaubercalculation in zero rangeapproximation (relative S→ δ(r12)) agrees as well

I Last time: very softA-dependence


EMC vs SRC

Experimental Observation: EMC vs SRC



L.B. Weinstein et al. PRL106 052301 ('11)O. Hen et al. PRC85 047301 ('12)

I EMC effect quantified by slopeof ratio F2A/F2N .

I Recent observation that a2(A)and EMC slope show a linearcorrelation.

I Suggests that both phenomenamight be driven by local densityand/or high virtualityfluctuations.

I Several experiments planned forJLAB12 (deuteron EMC, isospindependence, more a2, EMCmeasurements,...)



Number of relative S-pairs pernucleon shows linear correlationwith EMC slopes.

I EMC effect quantified by slopeof ratio F2A/F2N .

I Recent observation that a2(A)and EMC slope show a linearcorrelation.

I Suggests that both phenomenamight be driven by local densityand/or high virtualityfluctuations.

I Several experiments planned forJLAB12 (deuteron EMC, isospindependence, more a2, EMCmeasurements,...)


Conclusions

I The number of SRC-prone pairs in a nucleus A(N,Z ) is proportionalwith the number of IPM pairs in a relative S state (n = 0,L = 0) orclose together

I The number of SRC-prone pairs follows a robust power law forpp,pn,nn: ∼ A1.44±0.01

I Inclusive A(e, e ′) at 1.5 . xB (2N): The a2 (A/D) can be predicted andthese predictions are not inconsistent with trends and magnitude of thedata (corrections for c.m. motion)

I Inclusive A(e, e ′) at 2.2 . xB (3N): Fair prediction for the a3(A/3He

)

I Exclusive A(e, e ′pp): scaling behavior of cross section with c.o.m. momentum ∼ F (P):

Confirmed! very soft mass dependence of cross section: Confirmed!

I Magnitude of the EMC effect −dREMCdxB

is proportional with thepredicted number of SRC-prone pairsWim Cosyn (UGent) JLab Physics Division Seminar August 16, 2013 30 / 30

Documents

Mass dependence and scaling properties of nuclear short- range … · 2013-08-21 · Mass dependence and scaling properties of nuclear short- range correlations Wim Cosyn, Camille